Split testing is a technique you can use to test which variation performs better. It’s very common in web design. An online store may want to find which color checkout button converts the best. They would then show half their audience one colour, and the rest another and track conversion rates. It’s very important to split test your Amazon listing to help optimise it.
Andrew Browne co-founded Splitly as a way to use this technique with Amazon listings. This way you are able to test different aspects of your listing, to fine-tune every feature so it converts as good as it can. You can test different main images and title to find out which gets better traction. This uses objective data rather than your subjected intuition.
It is the first thing your customers see therefore has a major impact on getting potential buyers’ attention. If you split test your Amazon listing image, you can determine the best picture that will get people to click on your product.
One success story that Andrew shared was one he did on his own product. He wasn’t looking for results, he was merely testing the software. In doing so, he found that by changing the main image, to one he thought were do poorly, he improved his conversion rate by 50%.
Another great place to start is your price. It’s easy to test and is incredibly impactful on your conversion and you can get your results quickly. Other places to test are your title, bullets, and product description.
Splitly is an amazing tool that does the split testing for you. When your start you test, your listing updates at midnight, Amazon time, which is GMT-8. The reason for that is Amazon does your reports based on that time. Then every night at midnight, you listing updates with the other version of your listing. While it’s not true AB testing, it’s the best you can do with Amazon’s limitations, and it’s better than using your own intuition.
It’s important to understand that split testing isn’t for saving products. If you have a failing product, it won’t help to split test to try and save it. You need to have enough traffic in order to get enough data to make improvements. You want to split test products that are already doing well in order to improve them and boost them a few percent. You’re looking to improve upon products that are already getting traffic and sales.
How long it takes to split test a listing depends on a few factors. The more sales your product gets, the quicker you’ll get your results. If you’re selling 5 units a day, Splitly will be able to collect enough data much quicker than if you’re selling 1 unit a day.
The significance of the difference places a large factor. If you’re testing a price difference of 5 cents, it might not affect the sales enough to determine a winner. Whereas if you’re testing a $5 difference, it will have a larger effect and the software will be able to determine which is better.
There is no specific number, it just depends on how long it takes Splitly to gather enough data to give you an answer. It may determine that there isn’t enough difference to go one way or another.